THE MICROFOUNDATIONS OF POLITICAL CLIENTELISM: LESSONS FROM THE ARGENTINE CASE

Mariela Szwarcberg

Working Paper # 377 – July 2011

Mariela Szwarcberg, a postdoctoral lecturer at the University of Chicago’s Center for Latin American Studies, conducts research in several areas related to the consolidation of democracy in Latin America. Previously, she was a visiting fellow at the University of Notre Dame’s Kellogg Institute for International Studies and at Yale University’s Program on Democracy and has served as a consultant to the National Institute for Public Administration in Argentina. Szwarcberg holds a PhD in political science from the University of Chicago.

ABSTRACT

This paper challenges the assumption that parties and candidates with access to material benefits will always distribute goods to low-income voters in exchange for electoral support. I claim that a candidate’s capacity to turn to clientelistic strategies of mobilization is a necessary but insufficient condition to explain his or her decision to use clientelism. Besides having access to material resources and a network of party activists to distribute goods to potential voters, candidates have to prefer to use clientelism to mobilize voters. In studying candidates’ capacities and preferences to use clientelism, the paper provides an account of the microfoundations of political clientelism in Argentina. By combining quantitative and qualitative data at the municipal level, I find that the number of pragmatist candidates, who are capable of using clientelism and prefer to turn to these strategies, is almost equaled by the number of idealist candidates who, although capable, prefer not to use clientelism.

RESUMEN

Este estudio plantea la hipótesis de que la capacidad de los candidatos de utilizar estrategias clientelares es una condición necesaria pero insuficiente para explicar su decisión de distribuir bienes y beneficios materiales a cambio de apoyo electoral. Mas allá de tener acceso a bienes materiales y redes de activistas políticos que colaboren con la distribución de bienes, los candidatos deben también preferir usar estrategias clientelares. Este artículo combina datos cuantitativos y cualitativos a nivel municipal en Argentina para mostrar que no hay una diferencia significativa entre las preferencias de los candidatos que cuentan con la capacidad de usar clientelismo. El artículo encuentra una cantidad similar de candidatos que prefieren utilizar estrategias clientelares, candidatos pragmáticos, y candidatos que prefieren rechazar el uso del clientelismo, candidatos idealistas.

Szwarcberg

1

In her seminal article about machine politics in Argentina, Susan Stokes (2005) describes

how party operatives are forced to make choices about how to distribute limited

resources.

Machine operatives everywhere face a version of the dilemma that an Argentine Peronist explains. About 40 voters live in her neighborhood and her responsibility is to get them to the polls and get them to vote for her party. But the party gives her only 10 bags of food to distribute, “ten little bags,” she laments, “nothing more.” (Stokes 2005, 315)

Stokes concludes that the party operative will give the bags of food to those swing

voters who will support her party only in exchange for a bag. Stokes also argues that the

operative will monitor those voters who receive the bags to make sure that they hold up

their end of the clientelistic deal. Building on Stokes, Simeon Nichter (2008) argues that,

given the constraints of the secret ballot, party operatives will monitor voter turnout

instead of vote choice because monitoring electoral participation simply requires

observing whether the voters who have received bags of food went to vote.

A new research agenda is focusing on modeling formally and testing empirically

how clientelistic parties will distribute goods to maximize vote returns (Díaz-Cayeros,

Estévez, and Magaloni 2008; Dunning and Stokes 2009; Gans-Morse, Mazzuca, and

Nichter 2009; Rosas and Hawkins 2007; Kitschelt and Kselman 2010). While this

literature undoubtedly enhances our understanding of the dynamics of distributive

politics, it assumes that party operatives, such as the one described by Stokes, will always

distribute goods in exchange for support. However, party operatives in Argentina and

elsewhere can prefer not to distribute goods in exchange for electoral support, thus

forgoing the use of clientelistic strategies or clientelism (terms that I use

interchangeably), as the testimony of an Argentine Peronist operative in Buenos Aires

highlights:

I’ll never get 20 bags of food, drive to a neighborhood, and say: “This is for you to vote for me.” I’ll give voters the bags: “Take this because you need it. Chau!”1

This paper questions the assumption that parties and party operatives with access

to material benefits such as bags of food will always distribute goods to low-income and

Szwarcberg

2

working class voters in exchange for electoral support. Instead, I argue that a party

operative’s capacity to turn to clientelistic strategies of mobilization is a necessary but

insufficient condition to explain the use of clientelism. Besides having access to material

goods and networks of activists, party operatives have to prefer to build clientelistic

linkages with voters.

In advancing a distinction between candidates’ capacities and preferences to use

clientelism, this paper aims to improve existing theories about and measurement of

clientelistic and programmatic linkages between politicians and voters in democracy

(Kitschelt and Wilkinson 2007). To examine the effects of a set of incentives on

individual party operative preferences, I used original evidence from municipal

candidates in Argentina. The country’s institutionalized party system (Mainwaring and

Scully 1995) enables me to test whether and how often candidates affiliated to stable

parties with roots in society and solid party organizations choose to use clientelism to

mobilize voters.2

Argentina uses a system of proportional representation with closed-list ballots in

which a candidate’s position on the party ticket determines his or her chance of being

elected. Party mayors decide a candidate’s position on the ticket and thus, by distributing

positions, mayors are able to effectively reward or punish candidates based on their

ability to turn out voters. Consequently, party operatives interested in pursuing a political

career are encouraged to mobilize as many voters as possible to secure a higher-ranked

position on the ticket, which increases their likelihood of getting elected, reelected, or

promoted to a higher office. In focusing on municipal candidates, I was able to gather

systematic data for a large population of party operatives who vary in their capacity and

preference to use clientelism. Given the importance of voter mobilization in securing safe

list positions, why would any party candidate ever forgo the use of clientelistic strategies

that contribute to turning out voters?

Since 1983, Argentina has held free and fair elections with alternation in the

executive and considerable competition in provinces and municipalities. The two major

parties, the Radical Civic Union (Unión Cívica Radical, UCR) and the Justicialist

(Peronist) Party (Partido Justicialista, PJ), maintain territorial control over most

municipalities by combining a recollection of shared watershed historical events with

Szwarcberg

3

clientelistic inducements (Torre 2005; Auyero 2000; Levitsky 2003; Calvo and Murillo

2004; Szwarcberg 2009), creating “communities of fate” (Wellhofer 1979: 171) and

“electorates of belonging” (Panebianco 1988: 267). By making comparisons across

municipalities in two provinces with different political traditions, Buenos Aires and

Córdoba, I am able to test the effects of partisanship on candidates’ decisions whether or

not to use clientelistic strategies of mobilization.

CHOOSING STRATEGIES OF POLITICAL MOBILIZATION Candidates interested in pursuing a political career have to demonstrate their ability to get

votes for the party. The more votes a candidate manages to provide for the party, the

more likely he or she will be promoted. The testimony of Mario, a party candidate in

Buenos Aires, explains this logic sharply:

This is very simple. You are worth as much as the amount of people you can mobilize. You have a prize, a number. Your number is how many people you can carry to a rally and how many votes you can give in an election. I tell you, what you need to do is simple. How you do it, that is strategy.3

How candidates mobilize voters varies based on what they have to offer and what

potential voters need. In resource-rich, vote-poor precincts candidates build

programmatic linkages with voters, while in resource-poor, vote-rich precincts these

linkages are clientelistic. Candidates’ capacities to build clientelistic linkages are

determined by the combination of access to particularistic goods and their ability to

distribute these goods to those voters who are likely to turn out and support the party. As

a result, candidates affiliated to incumbent parties are more capable of distributing goods

for electoral support than candidates affiliated to opposition parties.

Current literature discusses the type of voter candidates target with particularistic

inducements: voters who are likely to support the candidate regardless of receiving

particularistic goods, core voters (Nichter 2008; Calvo and Murillo 2004; Ansolabehere

and Snyder 2003; Cox and McCubbins 1986), or voters who are likely to support the

candidate only if they receive a good in exchange, swing voters (Dahlberg and Johanssen

2002; Case 2001; Schady 2000; Nazareno, Brusco, and Stokes 2006; Dixit and

Szwarcberg

4

Londregan 1996). Within this framework, scholars are developing theories that comprise

different strategies targeted to different voters based on the strength of their partisanship

identification and propensity to turn out to vote (Dunning and Stokes 2009; Gans-Morse,

Mazzuca, and Nichter 2009). Yet, beyond a candidate’s capacity to use clientelism, he or

she has to prefer to use these strategies to mobilize voters.

Table 1 categorizes candidates according to their capacity and preference to use

clientelism. In the upper-left corner, one finds pragmatist candidates who are capable of

using clientelism and of employing these strategies to get promoted. In the lower-left

corner, there are resentful candidates who prefer to use clientelism but are unable to

employ these strategies. In the upper-right corner, one finds idealist candidates who are

capable of using clientelism but prefer not to employ these strategies, even if this

decision works against their interests in getting promoted within the party. Finally, in the

lower-right corner, one finds utopist candidates who have neither the capacity nor the

preference to distribute goods in exchange for electoral support.

TABLE 1

CANDIDATES’ CAPACITIES AND PREFERENCES TO EMPLOY CLIENTELISTIC STRATEGIES OF MOBILIZATION

Candidates’ preferences to employ clientelism to mobilize voters

Yes No

Candidates’ capacities to access and distribute clientelistic goods

Yes Pragmatist Idealist

No Resentful Utopist

Without observing an explicit choice to forgo clientelistic politics, we cannot be

sure whether candidates are resentful or utopist. In contrast, having the capacity to use

clientelism, idealist and pragmatist candidates make their preferences visible by deciding

whether or not to use clientelism.

Szwarcberg

5

CASE SELECTION AND DATA

Drawing on the comparative method, this paper combines quantitative and qualitative

observations of elected municipal candidates’ capacities and preferences to use

clientelism. The data were gathered between June 2005 and December 2006 in seven

municipalities in two Argentine provinces: Buenos Aires and Córdoba.4 In Buenos Aires,

I selected the municipalities of José C. Paz, San Miguel, and Bahía Blanca, and in

Córdoba, Río Cuarto, Villa María, Colonia Caroya, and the City of Córdoba.5 By

focusing on seven cases, I was able to carry out the extensive fieldwork that was

necessary to gather data on individual candidates’ capacities and preferences. The case

selection is based on the differences in population, housing quality, income, partisanship,

and incumbency that quantitative studies of vote buying and clientelism have used to

explain variation in parties’ selection of strategies of mobilization. Table 2 provides

sociodemographic and electoral information about the selected cases.

Descriptive statistics provide information about general patterns that, combined

with qualitative information, establish plausibility for the hypotheses proposed in this

paper. While the results presented here are confined to seven municipalities across two

Argentine provinces, I also interviewed local candidates and voters and attended rallies

and political meetings in other municipalities in Buenos Aires—Malvinas Argentinas,

Hurlingham, Avellaneda, Vicente Lopez, Quilmes, Merlo, La Matanza, Morón,

Ayacucho, and Pergamino. I also conducted fieldwork across other municipalities in the

province of Córdoba: Mina Clavero, Yacanto, Villa Carlos Paz, and San Francisco. In

2009, I did a follow-up field trip to Buenos Aires and the province of San Luis. The

information I collected in these districts supports the findings presented in this paper, and

thus I am confident that the selected municipalities are representative of a larger universe

of cases.

Szwarcberg

6

TABLE 2

SELECTED MUNICIPALITIES IN ARGENTINA

Municipality Population Number of low-income households

Number of social welfare beneficiaries (Plan Jefes)

Municipal incumbent in 2005

Córdoba Capital 1,272,334 369,793 50,389 New Party Río Cuarto 144,021 42,044 5,142 Radical Party Villa María 72,162 1,114 298 Peronist Party Colonia Caroya 13,806 4,018 211 Peronist Party José C. Paz 230,208 56,004 15,612 Peronist Party San Miguel 253,086 65,689 10,238 Peronist Party Bahía Blanca 284,776 88,260 5619 Peronist Party

Note: Population numbers are based on the 2000 national census (National Institute of Statistics and Census of Argentina, INDEC). The number of council members is legally stipulated and varies based on the population of each municipality. By combining educational, occupational, and construction characteristics, the INDEC measures the income levels of Argentine homes. A household that fulfills three of the following five characteristics is classified as low income: (1) a density per room that exceeds three inhabitants, (2) precarious physical conditions, (3) absence of indoor plumbing, (4) children aged between six and twelve years who do not attend school, and (5) more than four members per one employed member and a head of the household who has not finished primary school. Data about the most widespread welfare program, Plan Jefes, were collected by the author in each municipality for the year 2004. Incumbency describes the party in charge of the local executive in each municipality. In 2005, the Radical Party ran in alliance with the New Party under the label of More for Río Cuarto. I do not take into account the intraparty divisions within the Peronist party in the election and thus I do not distinguish between the Front for Victory and the Justicialist Party.

The sample comprises the 137 candidates who held elected positions as council

members in 2005. I traveled to the seven selected districts and conducted in-depth

interviews with the majority of the candidates who were mobilizing voters during a

national election in Buenos Aires in 2005 and a primary election in Córdoba in 2006. I

conducted fieldwork in Buenos Aires between June and December and was able to

participate in over 40 party rallies with the agrupaciones of the mayors of José C. Paz

and San Miguel. Agrupaciones are informal municipal Peronist organizations that cluster

neighborhood-level organizations (Unidades Básicas, UBs) in charge of “the bulk of the

party’s mobilizational work” (Levitsky 2003: 71). By focusing on the agrupaciones of

Szwarcberg

7

mayors in two municipalities, I was able to participate in every rally in which these

agrupaciones were present, either in their own or in neighboring municipalities, and I

was able to observe and talk with party candidates about their strategies to turn out

voters.6 By comparing the decision-making of candidates affiliated to the same party

competing for the same voters and with the same capacity to use clientelism, I have

gained invaluable knowledge of the variation in candidate preferences.

Half of the candidates in Buenos Aires were running to get elected or reelected, as

the local legislature in the province is renewed by halves every two years. In Córdoba, on

the other hand, candidates were campaigning to show their ability to turn out voters for

the party in order to be considered for reelection in the upcoming election of 2009. This

is because, in contrast to Buenos Aires, voters in Córdoba elect their mayors and

councilors together every four years.

Comparisons between the strategies pursued by candidates in Buenos Aires whose

tenure was going to be renewed in two years, and those who were running for reelection

and election in 2005 did not show dramatic differences. Neither did the strategies of

candidates who were on the top of the closed list, at the cutoff point where candidates

could either succeed or fail in getting elected, nor even below the cutoff point where

candidates were certain that they were not going to get elected. These findings reinforce

the argument advanced in this article that candidates have to constantly show their ability

to turn out voters to advance in their political careers.

To maximize the number of votes the party obtains, bosses distribute positions

based on how many voters each candidate is capable of turning out. Ballot positions

therefore reflect the value each candidate has for the party. Building on the argument

developed in this paper, one expects to find clientelistic candidates holding higher

positions than candidates who prefer not to use these strategies. However, several

candidates who hold middle and lower positions on the ticket also employ clientelistic

strategies, thus canceling out a significant effect. This does not reflect the inefficacy of

clientelistic strategies but, rather, differences in the length of candidates’ use of

clientelism. In fact, the data suggest that the longer candidates use clientelism to build a

constituency, the larger their following and the more likely they will climb to a higher

position on the party ticket.

Szwarcberg

8

I conducted 101 in-depth interviews and 36 semi-structured interviews with

elected candidates in their municipalities. In cases in which I could not interview the

candidates directly, I relied on information provided by key informants who were mostly

advisors who had known and worked for the candidates for several years, even decades,

and were thus able to provide knowledgeable and reliable information about candidate

preferences. The length of the interviews ranged from two to several hours, during which

candidates reflected about their decisions whether or not to use clientelism to get out

voters. As the qualitative section of this paper shows, candidates talked very openly about

their capacities and preferences.

TABLE 3

DATA GATHERED BY THE AUTHOR BETWEEN JUNE 2005 AND DECEMBER 2006

Municipality Number of council members

Number of in-depth

interviews

Number of semi-

structured interviews

Number of key

informants interviewed

Archival research (municipal level)

Córdoba Capital

31 20 11 5 La Voz del Interior La Mañana de Córdoba

Río Cuarto 19 15 4 2 El Puntal Villa María 12 9 3 3 La Voz del

Interior Colonia Caroya

7 5 2 2 La Voz del Interior La Mañana de Córdoba

José C. Paz 20 20 0 5 La Hoja

San Miguel 24 17 7 3 La Hoja Bahía Blanca

24 15 9 4 La Nueva Provincia

Total 137 101 36 24

I consulted the archives of La Hoja, a local independent daily newspaper that

focuses on the municipalities of San Miguel and José C. Paz, and reviewed La Nueva

Provincia for Bahía Blanca, El Puntal for Río Cuarto, and La Voz del Interior and La

Szwarcberg

9

Mañana de Córdoba for provincial information on Córdoba in order to provide external

validity for my participant observation research and ethnographic data.7 I carried out

additional archival research in national newspapers Clarín, La Nación, and Página/12.8

Besides providing descriptive statistics on the selected cases, Table 3 describes the

sources of information and the number and types of interviews conducted by the author in

each municipality.

CANDIDATES’ CAPACITIES TO USE CLIENTELISM

A candidate’s capacity to use clientelistic strategies of mobilization varies depending on

his or her access to resources and the existence of a network of party activists who

contribute to distributing those goods to voters who are likely to turn out and support the

candidate as a result of receiving particularistic goods. Incumbent candidates are more

likely to have access to material goods that enable them to solve voter problems than

candidates affiliated to opposition parties. Incumbents, however, have different levels—

national, provincial, and municipal—leading to different combinations in the access to

goods. Thus, a candidate who ran with a party that counted on the support of the

president (national incumbent), governor (provincial incumbent), and mayor (local

incumbent) would have more resources than one who could only count on the support of

the president. It is arguable that local support is as important as national and provincial

support given that municipalities can count on multiple resources to promote political

rallies and events and distribute goods to voters. Still, this paper focuses on candidates’

potential to use clientelism and not on the quantity of resources available to engage in

these strategies.

To distinguish candidates who were able to use clientelism from those that were

unable to employ these strategies I employ two necessary conditions. First, I distinguish

candidates who were affiliated to parties that held one or more executive offices in 2005.

Column 5 in table 3 provides descriptive statistics about candidates who ran with the

support of the local government (20.44 percent), national and local governments (13.87

percent), provincial and national governments (11.68 percent), and national, provincial,

and local governments (36.50 percent). Only 24 candidates (17.52 percent) did not have

any governmental support and, therefore, were unable to use clientelism.

Szwarcberg

10

Incumbent candidates could, nevertheless, be unable to distribute goods in

exchange for support. To be effective in using clientelism, candidates need not only

access to resources but also a network of party activists to distribute those goods to those

voters who are likely to turn out and support the party in exchange for receiving goods.

Building on this criterion, the second condition implies that only candidates affiliated to

parties that possessed an organization capable of distributing particularistic goods and

monitoring voters’ electoral support were able to employ clientelism.

In Argentina, only the Peronist (PJ) and Radical (UCR) parties have had

systematic access to public office and large networks of party activists capable of trading

favors for votes effectively. Scholars of Argentine politics have consistently highlighted

working and low-income voters loyalty to the Peronist party (Torre 2005; Levitsky 2003;

Calvo and Murillo 2004; Mora y Araujo 1995; Ostiguy 1998). Calvo and Murillo (2004)

show that political parties’ access to resources (supply side) and voters’ dependence on

fiscal largesse (demand side) benefit the Peronist party due to the geographic

concentration of its voters and its linkages with less skilled constituencies. In their study

of vote buying in Argentina, Brusco and her collaborators (2004: 70–71) show that voters

who receive a handout from a Peronist candidate are more likely to vote for the Peronist

party. The ethnographic and qualitative works of Javier Auyero (2000), Steven Levitsky

(2003), and Mariela Szwarcberg (2009) found further support for these arguments.

Table 4 shows that only 24 (17.52 percent) candidates were unable to use

clientelism: the remaining 113 candidates could turn to these strategies; and among those

62.04 percent were affiliated to the Peronist Party, 7.30 percent to the Radical Party, and

13.14 percent to other parties. Whereas Peronist candidates could count on the support of

the national, provincial, and local government in José C. Paz, Bahía Blanca, Villa María,

and Colonia Caroya, non-Peronist candidates could count on the support of the municipal

executive in Río Cuarto (UCR), and in the City of Córdoba (New Party).9 Luis Juez, a

former Peronist candidate and provincial anticorruption prosecutor, created the New

Party (Partido Nuevo, PN) to compete for office after being fired by the governor. The

name of Juez’s party summarized the leitmotiv of his political campaign: Córdoba needed

a change, something new, different from Peronism and Radicalism. Competing for votes

in a context where the party leadership of the PJ and the UCR was heavily questioned,

Szwarcberg

11

Juez ended Córdoba’s historical bipartisanship, becoming the mayor of the City of

Córdoba in 2003.10 The majority of the 18 elected candidates affiliated to the PN lacked a

network of activists and were thus unable to use clientelism. Yet, as I examine later,

candidates who participated in politics with either the Peronist or Radical Party before

joining this new party did possess networks of activists and the know-how to use

clientelism, and some of them indeed continued exchanging favors for votes as in the

past. Likewise, 10 Radical and 85 Peronist candidates had the possibility to engage in

clientelistic strategies of mobilization.

TABLE 4

INCUMBENCY AND PARTISANSHIP

Partisanship Total Incumbency Other

Party Radical Party

Peronist Party

Number of candidates

No incumbency 8 (5.84%)

16 (11.68%)

0 24 (17.52%)

Local incumbency 18 (13.14%)

10 (7.30%)

0 28 (20.44 %)

National and local incumbent

0 0 19 (13.87%)

19 (13.87 %)

Provincial and national incumbent

0 0 16 (11.68%)

16 (11.68 %)

National, provincial, and local incumbent

0 0 50 (36.50)

50 (36.50 %)

Total 26 (18.98%)

26 (18.98%)

85 (62.04%)

137

These findings suggest that Peronist candidates were more capable of using

clientelistic strategies than Radical candidates, but they do not provide information on

candidate preferences. Are Peronist candidates more likely to prefer to use clientelism

than Radical candidates under similar circumstances? To answer this question I compare

the strategies employed by Peronist and Radical candidates who had the same capacity to

use clientelism when competing for the same voters.

Szwarcberg

12

CANDIDATES’ PREFERENCES TO USE CLIENTELISM

Having the capacity to use clientelism does not imply using it. Candidates have to also

prefer solving voter problems in exchange for electoral support. As the testimony from a

party operative in Buenos Aires quoted at the beginning of the paper illustrates, what

distinguishes clientelistic from not clientelistic candidates is not the use of resources to

solve voter problems, but the request that in exchange for solving problems, voters

support the candidate.

If I was using clientelism I would give voters bags of food only if they would vote for me, but I don’t do that, do you understand? I give them the bags because they need them. Of course, I will prefer them to vote for me, but if they need it, I’ll give it to them no matter what. Do I explain myself?11

Clientelistic candidates engage in solving voter problems to obtain their electoral

support and will thus monitor their participation. Without monitoring voter turnout,

candidates run the risk that voters will follow the political advice of opposition

candidates and “take the goods with one hand and vote with the other” (Szwarcberg

2004: 4). In countries where voting is compulsory, as in Argentina, and turnout numbers

are considerably high by international standards (Cantón and Jorrat 2003; IDEA 2009), it

is not possible to determine if voters turn out because they are mobilized, have strong

partisan preferences, or a combination of both. Indeed, Szwarcberg (2010) argues that

party bosses compare information from voter turnout at rallies and elections to judge a

candidate’s reliability and dole out rewards and punishments accordingly. Reliable

candidates who distribute goods to voters instead of pocketing them are rewarded with

higher-ranked positions on the closed list while unreliable candidates are punished with

lower-ranked positions.

By combining direct participation in over 40 rallies during the 2005 national

election in Buenos Aires, 5 rallies and a primary election in Córdoba in 2006, in-depth

and semi-structured interviews with candidates, and interviews with key informants, I

classified the mobilization strategies of 137 candidates. I consider that a candidate

engaged in clientelistic strategies of mobilization if he, she, or a designated party activist

took attendance of voter participation at rallies. To monitor voter participation at rallies,

Szwarcberg

13

candidates simply screen voters by taking attendance. Mabel, the private secretary of a

Peronist councilor in the City of Córdoba, explained to me that candidates use rosters—

“made in Excel and organized alphabetically”—of the names of beneficiaries of welfare

programs, public employees, and voters who had asked for favors. She said this while

showing the rosters she makes and updates “at least once a week, and during elections

almost daily.”12

Using attendance taken at rallies as a proxy to measure a candidate’s use of

clientelistic strategies enables me to discard a candidate who distributes goods to voters

without requesting their electoral support in exchange. Hence, a candidate who does not

monitor voter participation at rallies is not classified as clientelistic. Still, it is possible

that a candidate monitors voter participation at rallies and not at elections and vice versa.

Building on Szwarcberg (2010), I expect clientelistic candidates to prefer monitoring

voter participation at rallies rather than at elections because rally performance is easier to

measure and reward than voter turnout at elections. Szwarcberg’s study of voter turnout

in Argentina argues that clientelistic candidates tend to rely on clientelism to mobilize

voters to participate in both, rallies and elections, and monitor both political events

accordingly.

Table 5 shows that, although not even, the division between candidates who

distributed goods in exchange for participation and those who did not engage in these

strategies to turn out voters was very uniform: 63 candidates used clientelism (45.99

percent), and 74 candidates (54.01 percent) did not. Among those who engaged in

clientelistic strategies, 52 candidates were affiliated to the Peronist Party, 7 to the Radical

Party, and 4 candidates were former Peronist candidates who had ties to networks of

party activists now affiliated to the New Party. Most candidates affiliated to the New

Party preferred not to use clientelism, but these 4 candidates, nevertheless, chose

otherwise. I examine these cases in detail in the following section.

Among candidates who did not use clientelism, 22 were affiliated to other parties,

19 to the UCR, and 33 to the PJ. To explain why 52 candidates affiliated to majority

parties with the capacity to use clientelism preferred not to turn to these strategies, I

combine the use of descriptive statistics with life histories and in-depth and semi-

Szwarcberg

14

structured interviews. I also examine the preferences of candidates affiliated to other

parties who opted out of pursuing clientelistic strategies.

TABLE 5

PARTISANSHIP AND CLIENTELISM

Candidates’ partisanship

Other Party

Radical Party

Peronist Party Total

Did the candidate take attendance of voter participation at rallies?

No 22 (16.06%)

19 (13.87%)

33 (24.09%)

74 (54.01%)

Yes 4 (2.92%)

7 (5.11%)

52 (36.96%)

63 (45.99%)

Total 26 (18.98%)

26 (18.98%)

85 (62.04%)

137 (100%)

EXPLAINING VARIATION IN PARTY CANDIDATES’ SELECTION OF STRATEGIES OF POLITICAL MOBILIZATION

My theory predicts the existence of four types of candidates based on the combination of

their capacities and preferences to use clientelism. Table 6 uses comparative data from

Argentina to categorize candidates based on this schema, showing not only the existence

of idealist candidates but also that the number of pragmatist (59 candidates) and idealist

(52 candidates) was almost even. In failing to consider that 52 out of 111 candidates

prefer not to use clientelism, the literature both miscalculates the extent of clientelism and

misinterprets candidate preferences. First, in ignoring candidate preferences, the literature

assumes that in these cases 111 candidates will use clientelistic strategies when only 59

actually did employ clientelism to turn out voters. Second, in making policy makers

aware of both the existence and the significant numbers of candidates who prefer not to

use clientelism, the current work will help to make a successful case for designing

institutional incentives that will promote the political careers of idealist candidates.

Szwarcberg

15

TABLE 6

MUNICIPAL CANDIDATES’ CAPACITIES AND PREFERENCES

Candidates’ preferences to employ clientelism to mobilize voters

Yes No Total

Candidates’ capacities to access and distribute clientelistic goods

Yes Pragmatist 59

(43.07%)

Idealist 52

(37.96%)

111

(81.02%)

No Resentful

4 (2.92%)

Utopist 22

(16.06%)

26

(18.98%) Total 63

(45.99%) 74

(54.01%) 137

(100%)

Candidate testimonies collected in this paper highlight the importance of having

access to material resources to turn out voters and how this capacity induces candidates to

prefer to use clientelism.

Money is fundamental. If you don’t have money you can’t do anything in politics: You can’t solve voter problems, you can’t mobilize people (no podes tener gente).13

In linking having access to material resources with the possibility of solving voter

problems, candidates’ partisanship affiliation becomes a key variable in explaining

variation in candidates’ preferences. To study partisanship effects in candidates’

preferences, I compare cases of candidates affiliated to the same party running in the

same election and under similar circumstances. Minority parties’ lack of access to

government resources mostly prevents them from turning to clientelistic strategies of

mobilization, and thus I focus on the cases of the PJ and the UCR. I also study the unique

case of the PN as it succeeded in winning a local election by using programmatic

linkages with voters.

By making comparisons among candidates affiliated to the PJ in José C. Paz, to

the UCR in Río Cuarto, and to the PN in the City of Córdoba, I was able to conduct in-

depth interviews with every elected candidate while participating in several political

Szwarcberg

16

mobilizations and activities. In comparing the strategies chosen by candidates affiliated to

the same party, competing for the same voters, and with the same capacity to use

clientelism, I was able to hold variables such as age, gender, education, and income, as

well as capacity to use clientelism, constant and focus on variation on candidates’

preferences, thus gaining internal validity for my argument.

One of the poorest municipalities in Buenos Aires, José C. Paz could be easily

classified as a giant shantytown. Socio-economic indicators collected during the 2001

national census find that 44 percent of the municipality households live in precarious

homes and 7.7 percent of those households experience situations of critical overcrowding

(hacinamiento crítico). More than a quarter of the municipality’s 230,208 residents live

without meeting their basic needs (necesidades básicas insatisfechas, NBI) such as

indoor plumbing, employment, and education, and 63.2 percent of the residents do not

have health insurance. More than half of the population have not finished high school and

fewer than 10 percent attended college.

The municipality has 160 soup kitchens and 6,400 unpaved roads (calles de

tierra). Local authorities delivered between 4,000 and 5,000 bags of food daily during the

economic crisis in 2001, and almost 50 percent of the population received state aid,

mostly in the form of welfare programs. During that time, 80 percent of the economically

active population was unemployed.14 Under these conditions, incumbent candidates could

easily mobilize voters by simply distributing bags of food, a strategy that several of them

pursued. In explaining or justifying their decisions, pragmatist candidates referred to a

more or less explicit conception of realpolitik, defined as a system of politics or

principles based on practical rather than moral considerations, to explain their decision to

use clientelism.

The only thing to eat is shit and there isn’t enough for everyone. Under these circumstances, one cannot think of an ideal world. Either you go home or you stay in a coffee shop philosophizing about how it should be. It is messed up, but the rest of the reasoning is immature in that it confuses what should be with what is. The activists have to stay true to their principles everyday. Yeah, that is wonderful, you know, but if I think like that I’m a romantic without practical consequences.15

Szwarcberg

17

The dynamics of intra-party competition induces candidates to engage in

clientelism because if they do not exchange goods for support, someone else from their

party will and get their political promotion instead.

Candidate: In that election, they bought our voters in the bus we had rented to mobilize people. Author: Dirty? Candidate: Dirty or not, that’s politics. You can be the Mother Teresa of Calcutta, but in politics you don’t go anywhere if you don’t know how to play these games. When the definite moment comes, you have to show what you got (poner la carne en la parrilla). There is always someone who believes that he is better than you are, and they are convinced that you are trash, that you’re completely unworthy. There is always someone competing with you, ready to cut your throat (serruchándote el piso), because he wants to be in your position. And instead of being happy because you had been elected, he wants to be elected even though he doesn’t have the capacity to be an elected official (aunque no le de el cuero). The worst among politicians are never those who are in front of you but those who are by your side. The ones who are in front of you compete against you by using another image, with another program. But, those who are supposed to be with you, those are the worst of all.16

Still, there are idealist candidates who, having the possibility to use clientelism,

prefer not to use these strategies even when this will mean committing political suicide.

Candidates who prefer not to use clientelism fail not only in mobilizing voters but also in

sending party leaders the signal that they are willing to do what it takes to remain in

power. Idealist candidates are neither naïve nor inept; they understand well how

clientelism works, yet they prefer not to use the strategies that would secure their tenure

in office. Peronist idealist candidates in José C. Paz as well as Radical idealist candidates

in Río Cuarto shared the idea that political action should be guided by a normative

commitment to social justice.

By comparing the preferences of candidates competing for the same voters under

the same conditions in José C. Paz, I am able to examine the causes that explain why

some candidates prefer to use clientelism while others prefer to avoid engaging in these

strategies. Out of the 20 elected candidates in José C. Paz, all but one were affiliated to

Szwarcberg

18

the Peronist party. Sergio Formenti, a candidate affiliated to the Federalist Union Party

(Partido Unidad Federalista, PAUFE) is examined below as a representative of a

resentful candidate.17 Here, I focus on three Peronist candidates affiliated to the mayor’s

agrupación who worked in low-income neighborhoods that are comparable in terms of

size, voter partisanship, and propensity to turn out. I chose these cases for two reasons.

First, these three enable me to learn about candidates’ preferences, given that they

represent cases of candidates who made different choices with regard to using clientelism

despite facing the same opportunities. Second, two out of the three candidates preferred

not to use clientelism, providing me with an opportunity to understand the reasons why

candidates opt to commit political suicide.

Néstor Solis exemplifies the mind of idealist candidates who, while being aware

of committing political suicide by rejecting the use of clientelism, prefer, nevertheless, to

pursue this path. Candidates from his own and rival parties, key informants, and party

strategists agreed on highlighting Solis’s political potential. In the words of one activist:

“he could have been reelected easily (with emphasis) if he was a little bit more

flexible.”18 By flexibility, this activist was referring to Solis’s well-known rejection of the

use of clientelism to get political support. When I asked Solis why he rejected using a

strategy that he knew would allow him to get reelected and thus conduct politics on his

own terms, he replied:

I just don’t believe in a clientelistic political construction. It’s that simple. I believe that supporters have to choose to become part of a political project after discussing ideas, policies, not salaries.19

Although he obtained a significant number of votes, Solis failed to get reelected

when competing against the political machine of José C. Paz mayor, who had a personal

grudge against him and thus deployed additional money to make sure that Solis’s low-

income supporters had a hard time getting to the polls to support his candidacy.20 At the

time of the interview, Solis was selling acrylic paint while still participating in afternoon

political meetings in his neighborhood. In these meetings voters discussed political issues

such as: Who should be taxed in the municipality? Who should have a right to receive

state aid? How should that aid be distributed to guarantee that voters take those goods as

rights and not as political favors with conditions attached? These political meetings were

Szwarcberg

19

significantly different from the meetings of other Peronist agrupaciones in the

municipality, where the majority of the attendants were there because otherwise their

benefits would have been taken away and there was no discussion of politics but, rather,

of logistics: Who is going to mobilize voters in each neighborhood? Who is going to go

house-by-house to inform voters about an upcoming party rally?

Neighboring candidate Juan Carlos Romano, whose tenure was going to expire in

two years, echoed Solis’s preferences as well as his awareness of the consequences of the

rejection of the use of clientelism. During our conversations before the 2005 election,

Romano acknowledged that if he wanted to boost his turnout numbers, he had to threaten

voters who were receiving benefits and those whose problems he had solved because,

otherwise, it was very likely that those voters would go to support another candidate from

his party.

Voters are not bad people, they just have dire needs, and so they will support whomever helps them to solve their problems. It’s simple, even understandable, and straightforward.21

Romano also informed me about the effects that his preference not to use

clientelism had on candidates from his party who made the opposite choice. “They don’t

forgive me,” he told me, as his tone of voice began changing and his eyes got wet.

Author: You got emotional… Romano: No. It’s that I have my ideals, right? And there are things that I will not accept because they go against what I think. And I say it, I make it manifest.22

It was a strange moment, and he asked me not to keep on talking about this issue.

There was a long silence and I felt sympathy for him. Romano was raised embedded in a

popular machista culture where men do not cry at all, much less in public and in front of

a woman. Romano’s tears capture the impotence a candidate feels when choosing not to

use clientelism. At the end of our interview, Romano conceded: “this is how politics

work.” Yet, understanding how politics work is also what explains why some candidates

prefer to use clientelism, as the case of pragmatist candidate José Mondovi illustrates.

Mondovi envisions politics as a boxing match without referees.

Szwarcberg

20

I get up to the ring to box with gloves, but if you kick me in the knee, I’ll kick you back. If I don’t kick you, I’ll lose, you’ll win, and there won’t be a judge to tell me: “You Sir are correct.” You are kicked out, and left alone, crying, and that’s a pretty thing about politics. There are no untouchables. No one will look for you. No matter how much they respect you, and tell you that you are great. No one is going to make an effort for you. This isn’t bad; it’s just the rules. One cannot take things too personally but must use the rules of the game Either you get used to it, or you go crazy, or you leave.23

In José C. Paz, Peronist and not-Peronist candidates observed that those who

engage in clientelism succeed in their political careers without being effectively

penalized by either the party or the courts. As a result, candidates interested in pursuing a

political career who have the capacity to employ clientelistic strategies are strongly

encouraged to turn to these strategies.

Radical candidates in Río Cuarto, a municipality that only once since the return of

democracy had a Peronist government, experienced the tension between an idealist and

pragmatist campaign during the primary of the Radical Party in 2000. At the time, two

former reelected mayors, Miguel Angel Abella and Antonio Benigno Rins, took different

approaches to recover the local administration for their party: while Rins favored building

an electoral alliance with the PN to secure the UCR’s electoral victory, Abella openly

rejected such a strategy and campaigned against compromising the UCR’s principles by

building pragmatic alliances. It was a highly contested primary, which Rins won to

become again the mayor of Río Cuarto by joining forces with More for Río Cuarto (Más

por Río Cuarto, MRC) the electoral coalition between the UCR and the PN.

When I asked Abella what he had learned from that experience he replied that he

still held the same convictions and would have done exactly the same no matter what. I

asked him if he didn’t feel like David fighting against Goliath for trying to compete using

ideals, to which he replied that he envisioned himself rather as Moses walking in the

desert—he and other idealist activists believed that they were training a new generation

of politicians who would eventually get elected based on programmatic and not

clientelistic appeals.

Szwarcberg

21

Utopist candidates’ beliefs mirrored those of idealist candidates, but their inability

to distribute goods makes them ineligible to commit political suicide. Opposition

candidates affiliated to parties with limited resources are unable to turn to clientelism. In

practice, this means that neither resentful nor utopist candidates will employ clientelistic

strategies, but for significantly different reasons. Resentful candidates are unable to use

clientelism because they do not have access to material goods, while utopist candidates

would not use clientelism even if they had that access. The cases of resentful and utopist

candidates comprise 19 percent of the sample, and it is worth noting that there were 13.14

percent more utopist than resentful candidates.

Resentful candidates constantly referred to what they define as unfair

competition, that is, competing to mobilize low-income voters in situations in which

some candidates—those affiliated with incumbent parties—have more access to

resources than other candidates—those affiliated to non-incumbent parties.

In reality, there is less conspiracy than it seems. For instance, in the 2001 election, an election of which we are very proud, I went to the neighborhoods of very poor people, people with whom I had worked a lot, people who knew me and liked me, and nevertheless, one of the guys who had helped me came to ask me for money because he needed to buy some construction material to repair the roof of his house. He told me that he was not asking for money to vote for me, he said that it was OK if I didn’t have money, but he said that I had to understand that people needed money. Today you can’t mobilize 20 voters if you don’t buy them.24

This was the reasoning of Sergio Formenti, a candidate affiliated to the PAUFE

who was competing to mobilize the same voters targeted by Peronist candidates in José

C. Paz, a municipality where Peronist candidates could count on the support of the

national, provincial, and municipal governments. Following Formenti’s reasoning, if he

had been able to distribute construction materials, he would have been able to get voters’

support. Candidates like Formenti, who were unable to solve voter problems, constantly

pointed out their incapacity as the reason for their failure to mobilize voters.

Unfortunately, voters listen to you, they are interested in you, but they need things. Then, if you do not have money, if you can’t give them things, they can’t support you. They

Szwarcberg

22

support whoever has things to give away, no matter who she or he is.25

Table 7 describes the partisanship affiliation of each type of candidate further

demonstrating the strong relationship between Peronism and clientelism, since the

majority of pragmatist candidates were affiliated to the PJ. Still, the number of idealist

candidates within the Peronist party was significant. Partisanship differences also

highlight the fact that Radical candidates were less prone to use clientelism than Peronist

candidates. The high number of utopist candidates is driven mostly by the emergence and

success of the New Party in Córdoba. Four out of 18 candidates affiliated to the PN did

not distribute particularistic inducements to voters. The majority of the New Party

candidates were successful businessmen, professionals, and professors who decided to

participate in politics for the first time and thus had neither the know-how nor the

networks of activists that would have enabled them to use clientelism. Indeed, it is

plausible to suppose that Juez selected these candidates precisely because they were new

to the existing political establishment and thus unfamiliar with the old clientelistic

strategies used to turn out voters.

TABLE 7

CANDIDATES’ CAPACITIES AND PREFERENCES AND PARTISANSHIP

Candidates Other Party Radical Party Peronist Party Total

Pragmatist 0 (0%)

7 (5.11%)

52 (37.96%)

59 (43.07%)

Idealist 0 (0%)

19 (13.87%)

33 (24.09%)

52 (37.96%)

Resentful 4 (2.92%)

0 (0%)

0 (0%)

4 (2.92%)

Utopist 22 (16.06%)

0 (0%)

0 (0%)

22 (16.06%)

Total 26 (18.98%)

26 (18.98%)

85 (62.04%)

137 (100%)

In contrast, the four candidates who resorted to clientelism were former Peronist

(three candidates) and Radical (one candidate) Party members. Placed in lower-ranked

Szwarcberg

23

positions on the party ticket, these candidates continued to mobilize voters by exchanging

favors for votes.

CONCLUSIONS

Political parties are organizations that seek to win elections, and to achieve this goal they

try to turn out as many voters as possible. By distributing rewards to candidates based

only on the number of voters they mobilized, parties encourage the use of clientelistic

strategies. Candidates who are capable of and prefer to use clientelism are encouraged to

employ these strategies by a perverse system of incentives that promotes the careers of

clientelistic candidates to the detriment of candidates who are either unable or unwilling

to use these strategies. Hence, it is not the case that candidates are always willing to use

clientelistic strategies but, rather, that those who refuse to engage in these practices are

unable to advance in politics.

When only candidates who use clientelism succeed in getting promoted within the

party, career-oriented candidates are indirectly, but successfully, encouraged to use

clientelistic strategies. In examining the causes and consequences that induce individual

candidates to prefer clientelism, this paper contributes to the literature by improving our

understanding of causal mechanisms. The perverse logic of incentives that induces

candidates to use clientelism becomes evident only if we focus on the conditions under

which candidates make decisions about how to mobilize voters.

A candidate who solves voter problems and does not mind if that voter is not

loyal to him is committing political suicide. Yet, the logic of an alternative, non-

clientelistic political construction is based on building trust with voters. Non-clientelistic

candidates tend to believe in the importance of building relationships of mutual trust and

respect between voters and candidates as the foundation of stronger and healthier

relationships of representation.

Szwarcberg

24

ENDNOTES

1 Interview conducted by the author in José C. Paz, Argentina, September 2005. This and all

subsequent translations from the Spanish are by the author. 2 Aside from methodological considerations, it is worth noting that some recent and seminal

studies of clientelism (Calvo and Murillo 2004), vote buying (Stokes 2005), and turnout buying

(Nichter 2008) are also based on empirical evidence from Argentina. 3 Interview conducted by the author in José C. Paz, November 2005. 4 Buenos Aires is the financial, productive, and political center of the country, whose voters have

the capacity to choose the country’s President, given the size of the province’s electorate. For

instance, José C. Paz, one of the municipalities examined here, has more than twice the number of

voters (120,000) as the entire province of Formosa (50,000). Córdoba is the third largest electoral

district after the province and the City of Buenos Aires. 5 José C. Paz and San Miguel are representative of the municipalities of the Conurbano. The

Conurbano comprises one-quarter (8,684,437 inhabitants) of the country’s total population in 1.2

percent of the territory and has the highest percentage of unemployed and illegally employed

workers. Bahía Blanca, in contrast, has a similar population to that of José C. Paz and San Miguel

but is located in the southern area of Buenos Aires. The City of Córdoba and Río Cuarto are the

two larger and most important municipalities in Córdoba, followed, among others, by Villa

María. Colonia Caroya is a small municipality located near the City of Córdoba. 6 Supporting material about the number, location (whether in the stronghold area of the province

or not), and participation (or lack of it) in each rally of the agrupaciones of the mayor of José C.

Paz and San Miguel is publicly available on the author’s website:

http://www.marielaszwarcberg.com/Mariela_Szwarcberg/Supporting_materials.html 7 I was also present at several of the interviews with party candidates conducted by La Hoja

journalists Alfredo Sayus and Fabián Domínguez. 8 I found further support for my results in works of investigative journalism (Verbitsky 1991;

Cerruti 1993; López Echagüe 2002; Vaca Narvaja 2001; O’Donnell 2005) that examine the

political trajectories of different national figures. 9 San Miguel’s mayor, Oscar Zilocchi, chose to work for the candidacy of Chiche Duhalde (PJ)

against the candidacy of the president’s wife, Cristina Fernandez de Kirchner (Front for Victory,

Frente para la Victoria, FPV). The Kirchner administration’s discourse of human rights directly

challenged Zilocchi’s party leader, Aldo Rico, who was involved in the military during the

Szwarcberg

25

country’s dirty war and had led a group of army mutineers, the caparintada movement, to rise up

against a recently elected democratic government. 10 The Peronist government of Germán Kammerath was such a debacle that party leaders knew voters would not pardon the party’s catastrophic administration in the City of Córdoba, and thus

the governor did not even campaign for the party in the most important district of the province.

With the PJ out of competition, the UCR, whose past administrations had been prized and

remembered by voters, could have benefited, regardless of the party’s national defeats. Yet, the

provincial and local party leadership was fractured. Eduardo Angeloz, who had governed the province between 1983 and 1995, had been charged with embezzlement; and although he was

found not guilty in 1998, there was still too much suspicion and discontent to nominate him

again. Ramón Mestre, his successor (1995–1999) and party rival, died in 2003. And finally,

Rubén Martí, who led the third faction of the UCR in Córdoba and was a former mayor of the city, was ill. Unable to nominate any leader of the party’s representative factions, the UCR

nominated Luis Molinari Romero, a qualified but uncharismatic candidate who was remembered

for being Angeloz’s right hand. In this regard, the party did not manage to fulfill the electorate’s

demand: a fresh face without ties to the past. 11 Interview conducted by the author in José C. Paz, October 2005. 12 Interview conducted by the author in the City of Córdoba, April 2006. 13 Interview conducted by the author in Villa María, April 2006. 14 Data collected from municipal authorities by the author. 15 Interview conducted by the author in San Miguel, November 2005. 16 Interview conducted by the author in José C. Paz, September 2005. 17 The PAUFE is a right-wing political party whose founder, leader, and former mayor of Escobar

(a municipality in Buenos Aires), Luis Patti, was police chief during the last dictatorship and has

been accused and convicted for torturing people during that period. 18 Interview conducted by the author in José C. Paz, November 2005. 19 Interview conducted by the author in José C. Paz, September 2005. 20 For instance, there is evidence that the mayor purposely restricted public transportation to the

neighborhoods that were likely to support Solis’s candidacy (La Hoja, 29 October 2005). 21 Interview conducted by the author in José C. Paz, August 2005. 22 Interview conducted by the author in José C. Paz, August 2005. 23 Interview conducted by the author in Bahía Blanca, September 2006. 24 Interview conducted by the author in José C. Paz, December 2005. 25 Interview conducted by the author in Río Cuarto, May 2006.

Szwarcberg

26

REFERENCES

Ansolabehere, Stephen, and James M. Snyder Jr. 2003. “Why Is There So Little Money in U.S. Politics?” Journal of Economic Perspectives 17 (1): 105–130.

Auyero, Javier. 2000. Poor People’s Politics. Durham, NC: Duke University Press.

Brusco, Valeria, Marcelo Nazareno, and Susan C. Stokes. 2004. “Vote Buying in Argentina.” Latin American Research Review 39: 66–88.

Calvo, Ernesto, and María Victoria Murillo. 2004. “Who Delivers? Partisan Clients in the

Argentine Electoral Market.” American Journal of Political Science 48 (4): 742–57.

Canton, Darío, and Jorge Raúl Jorrat. 2003. “Abstention in Argentina Presidential

Elections, 1983–1999.” Latin American Research Review 38 (3): 187–201. Case, Anne. 2001. “Election Goals and Income Distribution: Recent Evidence from

Albania.” European Economic Review 45: 405–23. Cerruti, Gabriela. 1993. El jefe, vida y obra de Carlos Saúl Menem. Buenos Aires:

Planeta. Cox, Gary W., and Matthew D. McCubbins. 1986. “Electoral Politics as a Redistributive

Game.” Journal of Politics 48 (2): 370–89. Dahlberg, Matz, and Eva Johansson. 2002. “On the Vote-Purchasing Behavior of

Incumbent Governments.” American Political Science Review 96 (1): 27–40. Díaz-Cayeros, Alberto, Federico Estévez, and Beatriz Magaloni. 2008. “Strategies of

Vote Buying: Poverty, Democracy and Social Transfers in Mexico.” Unpublished manuscript, Department of Political Science, Stanford University.

Dixit, Avinash, and John Londregan. 1996. “The Determinants of Success of Special

Interests in Redistributive Politics.” Journal of Politics 58 (4): 1132–55. Dunning, Thad, and Susan Stokes. 2009. “Clientelism as Persuasion and Mobilization.”

Unpublished manuscript, Yale University. Gans-Morse, Jordan, Sebastián Mazzuca, and Simeon Nichter. 2009. “Who Gets Bought?

Vote Buying, Turnout Buying, and Other Strategies.” Harvard University, Weatherhead Center for International Affairs, Working Paper 09-0006.

IDEA (International Institute for Democracy and Electoral Assistance). 2006. Voter

Turnout. http://www.idea.int/vt.

Szwarcberg

27

Kitschelt, Herbert, and Daniel M. Kselman. 2010. “The Organizational Foundations of

Democratic Accountability: Organizational Forms and the Choice of Electoral Linkage Strategy.” Unpublished manuscript.

Kitschelt, Herbert, and Steven Wilkinson. 2007. Patrons, Clients, and Policies: Patterns

of Democratic Accountability and Political Competition. New York: Cambridge University Press.

Levitsky, Steven. 2003. Transforming Labor-Based Parties in Latin America: Argentine

Peronism in Comparative Perspective. New York: Cambridge University Press. López Echagüe, Hernán. 2002. El Otro: Eduardo Duhalde: Una biografía política.

Buenos Aires: Grupo Editorial Norma. Mainwaring, Scott, and Timothy R. Scully, eds. 1995. Building Democratic Institutions:

Party Systems in Latin America. Stanford: CA: Stanford University Press. Mora y Araujo, Manuel. 1995. “Las bases estructurales del Peronismo.” In Manuel Mora

y Araujo and Ignacio Llorente, eds., El voto peronista: Ensayos de sociología electoral Argentina. Buenos Aires: Editorial Sudamericana.

Nazareno, Marcelo, Valeria Brusco, and Susan C. Stokes. 2006. “Réditos y peligros

electorales del gasto público y elecciones en Argentina.” Desarrollo Económico 46 (181): 63–86.

Nichter, Simeon. 2008. “Vote Buying or Turnout Buying? Machine Politics and the

Secret Ballot.” American Political Science Review 102 (1): 19–31. O’Donnell, María. 2005. El aparato: Los intendentes del conurbano y las cajas negras de

la política. Buenos Aires: Aguilar. Ostiguy, Pierre. 1998. “Peronism and Anti-Peronism: Class-Cultural Cleavages and

Political Identity in Argentina.” PhD Dissertation, Department of Political Science, University of California at Berkeley.

Panebianco, Angelo. 1988. Political Parties: Organization and Power. Cambridge, UK:

Cambridge University Press. Rosas, Guillermo, and Kirk Hawkins. 2007. “Turncoats, True Believers, and Turnout:

Machine Politics in an Australian Ballot System.” Unpublished manuscript, Duke University.

Schady, Norbert R. 2000. “The Political Economy of Expenditures by the Peruvian

Social Fund (FONCODES), 1991–95.” American Political Science Review 94 (2): 289–304.

Szwarcberg

28

Stokes, Susan C. 2005. “Perverse Accountability: A Formal Model of Machine Politics

with Evidence from Argentina.” American Political Science Review 99 (3): 315–25.

Szwarcberg, Mariela. 2004. “Feeding Political Loyalties: Clientelism in Argentina.”

Paper presented at the Comparative Politics Workshop, University of Chicago, IL. ———. 2009. “Making Local Democracy: Political Machines, Clientelism, and Social Networks

in Argentina.” PhD Dissertation, Political Science Department, University of Chicago, IL.

———. 2010. “Uncertainty, Political Clientelism, and Voter Turnout in Argentina: Why Parties

Conduct Rallies in Argentina.” Working Paper, Kellogg Institute for International Studies, University of Notre Dame, IN.

Torre, Juan Carlos. 2005. “Citizens versus Political Class: The Crisis of Partisan

Representation.” In Steven Levitsky and María Victoria Murillo, eds., Argentine Democracy: The Politics of Institutional Weakness, 165–80. University Park, PA.: Pennsylvania State University Press.

Vaca Narvaja, Hernán. 2001. El candidato: Biografía no autorizada de José Manuel de la Sota.

Buenos Aires, Argentina: Editorial Sudamericana. Verbitsky, Horacio. 1991. Robo para la corona: Los frutos prohibidos del árbol de la

corrupción. Buenos Aires: Planeta. Wellhofer, Spencer E. 1979. “Strategies for Party Organization and Voter Mobilization: Britain,

Norway, and Argentina.” Comparative Political Studies 12 (3): 169–204.